gamedev-GDC web games report 2004-03-31 - By NoiseCrime
Back
-- -- Original Message -- --
From: "Darrel Plant" <dplant@(protected)>
> Something that troubles me is that if it _is_ going to happen, it
> would have to be happening right now, and there haven't exactly been
> portents to that effect in the skies.
> Assuming that there ever is a major new version of 3D in the next
> release (eighteen months + slippage: ~Q4 2005), that'll be 4.5 years
> between updates of the 3D engine (D8.5 was released in early 2001).
> Shockwave will be 10 (announced in the fall of 1995, with developer
> tools released in December).
Well I guess I was optimsitically thinking of the next beta, or even an
incremental update, although I have to agree the points you make are valid
and have dampened my optimisnm somewhat. If it does stretch to 4.5 years
then you can bet that all other competitors will also have a new version by
then too, with performance improvements and a few new features. Its seems
sw3d is destined to fall by the wayside, and the longer its left without an
update the less likely or useful one will be. Its almost scandelous and
incompetent to allow possibly the largest browser penetration 3d solution to
fail due to lack of interest or motivation by the company who owns it (MM)
Then again so many other areas in Director seem to be reaching the end of
their productive life, sound for one is an absolute mess, judging by recent
posts. It may have once been acceptable not to get simulatnious sounds,
synched together but in this day and age its laughable. The text engine is
rapidly approaching a similar fate, its slow and awkard to use in some
cases, not really delivering all the features one would want, or in a
consistent manor. Quicktime suport appears to have lagged, and i can't
understand why we still don't have an image property for it. The list goes
on.
The annoying thing is that in a sense sw3d is so behind the times now, that
it needs such a radical overhaul to catch up, but in doing so I think its
highly unlikely that the engine would then need to be updated again for some
considerable time afterwards. Sure it will need bug fixes and tweaking but
little else, simply because there appears to be less developments happening
in 3D engines.
Undoubtably the biggest change has been the move to a programmable pipeline
(shaders), and this has given back a huge amount of flexability to the
developer. Whilst this is still likely to change, once a code path is in
place, there should be no reason why sw3d shaders couldn't keep up. Whilst
there are obviously some new tech developments to come such as moving to
more mathamtical surface defintions such as patches or nurbs, going for High
Dynamic Range Ligthing and the like. None of them I feel willl really be
must haves compared to the list of features we are currently missing such as
dynamic shadows (projection or stencil), bump mapping, vertex/pixel shaders,
normal mapping, texgen, proper RenderToTexture etc.
Obviously this is not an easy step for MM, and would require a considerable
investiment on their part. But then i think they'd have to accept that this
was always going to be an issue with going with a pre-built engine, that by
appearances can never be open sourced. If they had developed their own
propriety format or bought one like they did with Flash, then much of this
lag could have ben avoided with developers such as myself and many others
taking up the challenge to add new features.
Perhaps its time for someone else to step into the breach, and design a new
3d engine. Not that its an easy task, crossplatform, multiple API, and
mutliple card support is not for the faint hearted. Neither is adding bone
animation, SDS or LOD, but they are doable.
Perhaps a stepping stone solution is to go back and implement a basic openGL
interface. I was always disappointed that Ravware's version didn't get past
the beta. Obviously a striaght conversion of openGL functions to lingo is
not the best route to take. You need to add support for actual mesh
creation and loading of a file format, but thats not outside the realms of
possiblity, animation suport can be added through exisitng model formats
such as MD2 or MD3 and a bones player isn't too far fetched. From my
experince to date openGL is prety easy to get into and some open source xtra
might be viable.
Noisecrime 2004
__ ____ ____ ____ ____ ____ ____ ____ ____ ____
Dir3d-l mailing list
Dir3d-l@(protected)
http://nuttybar.drama.uga.edu/mailman/listinfo/dir3d-l
|
|
